Work Schedule Flexibility: A Contributor to Happiness?

LONNIE GOLDEN1
Penn State Abington, Pennsylvania, U.S.A

JULIA R. HENLY
University of Chicago, Illinois, U.S.A.

SUSAN LAMBERT
University of Chicago, Illinois, U.S.A.

Abstract
This article contributes to knowledge regarding determinants of happiness by examining the
independent role played by having discretion over one’s working time, using data pooled from
two years of a nationally representative US survey. Controlling for a worker’s income bracket
and work hours duration, having work schedule flexibility in the form of an ability to take time
off during the work day and, to a somewhat lesser extent, to vary starting and quitting times
daily, are both associated with greater happiness, whereas an ability to refuse overtime work is
weak at best. The associations are generally stronger among workers paid by the hour than by
salary. Worker utility functions thus may be enhanced by including the timing and flexibility of
working time. Policies and practices that promote more employee-centered flexible working
time may not only help workers alleviate work-life time conflicts, but also promote worker well-
being generally, especially among hourly-paid workers.

Keywords: Subjective Well-Being, Work Schedules, Workplace Flexibility, Working Time,

Economics of Happiness.

A growing body of literature on “happiness” explores empirically the complex influences of

income, inequality, hours of work and non-work time on individual as well as national well-
being.2 The conventional work-leisure model of utility suggests that greater current income will
be associated with higher levels of worker subjective well-being, even if produced by more hours
of work; however, this proves to be nuanced. In the present study, we aim to bridge this interest
among some economists in determinants of individual quality of life with parallel tracks in other
social science disciplines, which more directly examine work and workplace sources of well-
being.
With happiness now accepted in the realm of economics and discussions of national well-being,
the coast has been cleared for labor economics to apply available measures of happiness, and
study its relationship to various dimensions of working time (e.g., Di Tella & MacCulloch, 2006;
Stiglitz, Sen & Fitoussi, 2009; Clark, Frijters & Shields, 2008; Graham, 2009; Krueger, 2009;
MacKerron, 2011; Helliwell, Layard & Sachs, 2012).3 Non-pecuniary dimensions of well-being

1
Postal Address: 1600 Woodland Rd, Abington PA, USA. E-mail Address: Lmg5@psu.edu
2
Special thanks to Lauren G. Meadows and JaeSeung Kim for invaluable research support and to Peter Berg,
Heejung Chung, and Alex Herzog-Stein for comments.
3
Happiness research is part of a wider literature concerning dimensions of subjective well-being. It considers both
the experiential, affective state or daily life (hedonic) and the overall (global) life evaluation or appraisal varieties.

Electronic copy available at: http://ssrn.com/abstract=2129520

are considered important at the country level (Stiglitz, Sen & Fitoussi, 2009). A rich body of
empirical research has documented the potentially adverse effects at the individual worker level
of workplace inflexibility. Yet, discretion (i.e., employee-centered flexibility) heretofore has not
been included in economic studies of happiness or subjective well-being. Thus, it is important to
include not only the amount of work time or leisure (vacation or weekend) time (Kahneman &
Krueger, 2006; Hamermesh & Lee, 2007; Easterlin, 2010; Harter & Arora, 2010; Aaker, Rudd &
Mogilner, 2011), but also job and working conditions (Noll, 2011), such as the extent of
workers’ control over their timing of their work hours.
Our chief goal is to estimate empirically the relationship between a worker’s self-reported
happiness and his or her perceived discretion over the timing of work, controlling for both
income and the duration of working hours. We extend the basic economics of happiness
approach with elements of the job demand-control model (Karasek, 1979), by estimating the
extent to which a greater extent of worker control over working time matters, and may offset the
opposing negative effects of long work hours or lower income on worker well-being. We focus
on the relative size effects, in particular, of indicators of work schedule discretion, with special
attention to how relationships may vary for workers in hourly vis-à-vis salaried positions. This
research analyzes data pooled from two years of a nationally representative US survey that
permit observation of five indicators of working time – three assessing indicators of worker
discretion over work timing, plus two indicators of work duration. Multinomial regressions and
then Ordinary Least Squares regressions are conducted to test the extent to which an ability to
take time off during the work day, to vary starting and quitting times daily, and to refuse
mandatory overtime work are each associated with greater happiness, while controlling for the
influences of a worker’s income bracket and the hours duration of their work time per week (as
well as various demographic and other job characteristics). The size effects of these variables are
contrasted among hourly paid and salaried employees. Implications of the results for well-being,
public policy, and organizational practices regarding work time are then briefly explored.

The Income-Hours-Happiness Nexus as Background

Recent behavioral economics research tends to focus primarily on the issue of the amount of
absolute or relative income earned or leisure consumed. Subjective well-being, when measured
by self-reported happiness or life satisfaction, does not necessarily rise linearly nor
proportionally with increased income, particularly at higher levels of income, although the
estimated effects are mixed (Helliwell, 2003; Osberg & Sharpe, 2005; Kahnemann & Krueger,
2006; Binswanger, 2006; Clark, Frijters & Shields, 2008; Dolan, Peasgood & White, 2008;
Powdthavee, 2010; Zuzanek, 2012). The effect of income tends to dissipate, as people adjust to
their new absolute or relative income level, by adapting their aspirations or reference points
(Hagerty & Veenhoven, 2003). When the (unequal) distribution of happiness is decomposed, for
the US, there is a contributory role played by non-pecuniary factors in shaping it (Stevenson &
Wolfers, 2008). Happiness tends to depend on one’s family life experience, unemployment,
healthiness, and life cycle stage.4
The role of working time, including long work hours and the timing of work, has been largely
neglected in the happiness literature (Powdthavee, 2009; Spencer, 2011; van der Meer &

The paradigm of quantifying happiness, particularly into a single measure, is not without its philosophical detractors
(e.g., McCloskey, 2012), but also acceptors (e.g., Angner, 2010; Schubert, 2012; Hancock, 2013).
4
Also the cultural context, see Easterlin (2006), Mogilner, Kamvar, & Aaker (2011), and Frijters & Beatton (2012)
regarding the age and life-cycle patterns of happiness, and the legal context, see Huang (2010).

Electronic copy available at: http://ssrn.com/abstract=2129520

Wielers, 2013). Incomes are likely to be positively correlated with working hours, time spent
commuting to and from work and time spent away from family and friends. In contrast, time
spent at work, commuting and away from family and social lives are factors negatively
correlated with one’s global evaluation of life satisfaction or net affect (Kahneman & Krueger,
2006).5 Additional income from work, ex poste, may yield less than ex ante expected gains in
utility, to the extent that health and family life contribute separately from the effects of income.
Thus, omitting potential right-hand side variables may bias the estimated effect of income on life
satisfaction (Powdthavee, 2010).
Nevertheless, the omission of work hours duration, which tends to correlate positively with
income and negatively with elements of subjective well-being, does not seem to explain why
rising income has not produced commensurately rising happiness. Hours have declined over
time in advanced countries (other than in the US and Sweden), but without corresponding
increases in levels of happiness (Clark, 2005). Thus, the inclusion of hours actually magnifies the
puzzle of growing income without parallel rises in happiness (Clark, Frijters & Shields, 2008).
Moreover, subjective well-being is influenced less by the number of work hours per se than the
“mismatch” between workers’ actual and preferred work time duration (Sousa-Poza &
Henneberger, 2002; Bielinski, Bosch & Wagner, 2002; Tijdens, 2003; Van Emmerik & Sanders,
2005; Wooden, Warren & Drago, 2009; Drago, Wooden & Black, 2009). Curiously, many
individuals could conceivably reduce their own work hours without corresponding reductions in
their happiness, but they do not (Binswanger, 2006).6
Whatever the source, the income-happiness nexus is influenced by the income-work hours
relationship. It may eliminate entirely the impact of additional labor income on happiness
(Pouwels, Siegers & Vlaslom, 2008). The effect of longer hours may be more nuanced, however,
because the additional income it produces may increase job satisfaction, particularly for workers
with lower base income levels (Sousa-Poza & Sousa-Poza, 2000; Van Praag, Frijters & Ferrer-
Carbonell, 2003). Thus, while life satisfaction or happiness is clearly positively associated with
being employed, there may be an inverse-U relationship with hours duration. Longer work time
exerts a marginal disutility particularly for workers whose hours are already long, e.g., the
inverse association takes hold beyond 7 work hours per day (Knabe & Ratzel, 2010). Increasing
rates of involuntary part-time employment in European nations suggest that shorter hours alone
may not be a source of happiness for all workers (Mason & Salverda, 2009). Nevertheless, across
22 EU countries, workers with shorter or part-time hours are happier than full-timers, on
average. Importantly, this is largely because shorter-hour jobs provide flexibility to improve
current work-life balance.7
Much evidence collected from the fields of occupational health and organizational psychology,
labor-employment relations and work-life research suggests workers who work long hours often
experience reduced subjective well-being because they tend to experience added work stress,
fatigue and/or time conflicts, and thus face an additional risk of illness, injury, burnout or work-
to-family spillovers (e.g., Reynolds, 2003; Berg, Kalleberg & Appelbaum, 2003; White et al.,

5
For other externality effects of working time choices and scheduling, see Weiss (1996), Holzer (2005), Altman &
Golden (2007), Hamermesh & Slemrod (2008), Boushey (2011).
6
That may occur because individuals tend to underestimate the opportunity costs incurred with working for
additional income, and overestimate both the happiness yielded from increased income provided by additional work
hours and their future opportunities for time-saving (Binswanger, 2006).
7
e.g., Booth & van Ours, 2009; Okulicz-Kozaryn, 2011; van der Meer & Wielers, 2013. This has been formalized in
some countries’ government programs that reduce the cost of reduced hours by supplementing job earnings, such as
work-sharing policies and practices.
2003; Grönlund & Öun, 2010; Briscoe, Wardell & Sawyer, 2011).8 However, the relationship is
not always consistent between the volume of work hours per se and the subjective or objective
measures of life satisfaction (Hamermesh & Lee, 2007) or job satisfaction (Sousa-Poza & Sousa-
Poza, 2000; Green, 2004; Chongvilaivan & Powdthavee, 2012). Overtime work may involve
less risk if workers face only moderate overtime, favorable work conditions or attractive jobs
(Beckers et al. 2004), or if overtime is purely voluntary (Hamermesh, 1999; Friedman & Casner-
Lotto, 2003; Golden & Wiens-Tuers, 2006). However, when facing some supervisory pressure to
work overtime, the combination of involuntariness and working more than 50 hours a week has
been shown to elevate the risks of work-family interference, injury, illness, somatic stress,
depression and health complaints (Cornell Institute for Workplace Studies, 1999; Van Der Hulst
& Geurts, 2001). When the length, timing or scheduling of overtime work is not under the
control of the employee, any detrimental well-being effects of long hours of work tends to be
compounded or exacerbated (Ala-Mursula et al., 2002; Berg et al., 2004; Heisz & LaRochelle-
Côté, 2006; Beckers et al., 2008).

Discretion over Work Timing and Worker Well-Being

It may well be the incongruity between the desired and actual schedule of work hours that affects
one’s well-being, not just its duration (Krausz, Sagie & Bidermann, 2000). If the scheduling of
hours does not fit a worker’s preferred timing, individual welfare tends to be diminished
(Barnett, 2004). Moreover, more employee control over work time may moderate a negative
relationship between work hours and adverse outcomes (Berg, Kalleberg & Appelbaum, 2003;
Hughes & Parkes, 2007). For example, among those not permitted to change their own work
schedules toward their preferred schedules, 45 percent experience symptoms of “overwork,”
three times the rate among those who are permitted (Galinsky, Bond & Hill, 2005).
Flexitime work schedules, involving some employee input or discretion, often facilitate better
outcomes than traditional fixed-hour schedules. High-performance workplace practices and long
work hours, however, have interacted to trump positive effects of formal, work-life supports
(White et al., 2003). This may be in part because workers that have more flexible daily schedules
are also more likely to be working very long hours, perhaps as an act of reciprocation or
exchange (e.g., see Golden, 2009; Kelliher & Anderson, 2010). Moreover, when combined with
employee participation, schedule flexibility moderates adverse effects of longer hours, such as
work-life conflict (Wang, 2011). Having scheduling flexibility can sometimes be a double-edged
sword -- fixed work schedules might also offer workers a greater sense of predictability and
regularity (Schieman & Young, 2010; Lambert, Haley-Lock, & Henly, 2012).9 Nevertheless,
more flexible arrangements may improve employee attitudes and happiness, which in turn
increases worker performance (Atkinson & Hall, 2011). For example, greater scheduling
flexibility promotes lower sickness-absence, retention and customer satisfaction perhaps because
these workers are easier to work with and are happier (Robertson & Cooper, 2011).10 Control

8
Long hours of work per day or per week may undermine well-being not only through physical health (Brenner,
Fairris & Ruser, 2004; Dong, 2005; Beckers et al., 2008; Kattenbach, Demerouti & Nachreiner, 2010; Golden et al.,
2011), but also mental health (Virtanen, et al. 2012).
9
Having “flexibility” in the timing of one’s job is not necessarily equivalent to having “control” over one’s flow of
work (MacDermid & Tang, 2009).
10
In addition, flexible scheduling facilitates better coordination of daily work and life responsibilities (White et al.,
2003; Berg et al., 2004; Kelly, Moen & Tranby, 2011; Glauber, 2011; Kossek & Michel, 2011). Greater workplace
flexibility may indirectly promote worker productivity by improving workers’ health and healthy behaviors
(Galinsky, Bond & Hill, 2005; Grzywycz, Carlson & Shulkin, 2008; Loh, 2009).
over work schedules, including days off, is associated with reduced fatigue, sleep problems and
depression, which also promotes employee performance (Takahashi et al., 2011).11
In direct contrast to employee-centered flexibility, there are a myriad of reasons and motivations
for employers to make work hours more differentiated or variable to suit their own needs for
work hours and schedule adjustments (Bosch, 1999; Rubery et al., 2005; Bacon, Blyton &
Dastmalchian, 2005; Messenger, 2011; Lambert 2008). Greater variability in working time,
presumably due to employer-initiated adjustments, tends to reduce workers' well-being
(Askenazy, 2004), especially when it is irregular or unpredictable, and particularly at lower
incomes (Lambert, 2008). Conversely, lower variability in work hours is as important as higher
flexibility as a positive influence on meeting family and social commitments (Heisz &
LaRochelle-Cote, 2006; Costa, Sartori & Akerstedt, 2006; Olsen & Dahl, 2010). In sum,
whether framed as “discretion,” “control,” or “employee-centered flexibility” regarding the
scheduling of working time, there is reason to believe that its influence may extend beyond the
domain of work-life integration and into workers’ overall well-being and happiness.

Theoretical Foundation and Model

If work schedule flexibility over the timing contributes to worker well-being, including
happiness, then a standard utility (U) function ought to be amended with an additional argument,
beyond the conventional terms of income (Y) and leisure (L) time, – the degree of scheduling
discretion. This is captured by a coefficient (), in the expanded utility function expression:12

U = U [ Y ,L ;  ]
assuming: dU/dY,dU/dL, dU/d > 0.

Ultimately, or at least initially, it is employers that set work schedules (with the possible
exception of self-employment). Suppose there is an employer-set daily working time interval (I)
to which a given worker is scheduled. This may deviate from the worker’s preferred shift time
slot (denoted by I*). Worker well-being is improved with any increase in the degree of
responsiveness of their work schedules to changes in their preferred daily timing or interval (I*)
of work. Suboptimal utility occurs anytime actual schedules are slow to adjust toward (either
temporary or permanent) changes in a worker’s desired I*,13 as reflected by the term, .
Synchronization between a schedule required by employers, I , and the I* desired by employees
occurs only if the employer provides employees with a wider range of work schedule intervals
from which to choose, or an opportunity to self-adjust their own timing of work. Thus, even if a
worker is scheduled for his or her preferred duration of working time (e.g., an 8-hour workday),
the timing of the shift might create conflicts with other desired activities, such as caregiving or

11
Because at least some employees are prepared to accept in return for flexible work arrangements and place a
negative compensating differential, such as sacrificing a raise, bonus or alternative benefit (McCrate, 2005; Weeden,
2005; Heywood, Siebert & Wei, 2007; Winder, 2009; Felfe, 2012), they often carry a positive wage premium,
attributable probably to employee-centered flexible work facilitating a gain in productivity. Flexible timing of work
appears to boost average labor productivity and thus perhaps profitability in enterprises and industries (Shepard,
Clifton & Kruse, 1996; Gavin & Mason, 2004; Ortega, 2009; Lee & DeVoe, 2012) and perhaps also a consequent
improvement in technical efficiency for firms (Wolf & Beblo, 2004).
12
Tradeoffs between these three arguments may be subject to the usual concavity assumption.
13
For example, a worker might prefer an I* that starts at time (0) and finishes at time (n), e.g., 9am to 5pm or 7am to
3pm (under the simplifying assumption that workers prefer a single, continuous rather than split shift).
student work, or one’s natural circadian rhythms. An assumption of concavity suggests that those
individuals most deprived of work schedule discretion may gain relatively more from a given
increase in flexibility than a worker that already has much discretion. Similarly, workers whose
earnings are linked directly to the hours they work may gain the most from having schedule
discretion, because it may allow them to adjust the timing of their hours without reducing the
number of their hours and also their earnings.. An employer will presumably grant an employee
more scheduling discretion if the added short run cost of adopting flexible work scheduling is
exceeded by longer run labor cost savings or improved performance achieved by better matching
of actual to preferred schedules (Altman & Golden, 2007; Ortega, 2009; Lee & DeVoe, 2012).14

Study Hypotheses
In light of the existing literature and the implications of the above model, we postulate the
following four hypotheses regarding the relationship between happiness and aspects of hours
duration and work schedule discretion:
Hypothesis 1: When controlling for the potentially nuanced effects of higher income level,15 and
workers’ demographic characteristics, a longer duration of work hours (average weekly hours or
extra days of work per month) will have an ambiguous association with happiness.16 – It may be
negatively related to happiness because of adverse work to life spillovers and lost leisure time,
but also potentially positive to the extent that extra work brings additional current income (or
expected future reward) or non-pecuniary benefits.17
Hypothesis 2: Controlling for income, demographics, and duration of work hours, having greater
daily working time discretion will be unambiguously positively associated with happiness.18
Hypothesis 3: While all types of daily working time discretion contribute to well-being
positively, it is possible that some may have greater impact than others.
Hypothesis 4: For hourly paid and salaried workers, the association of happiness with work
schedule discretion may be similar in magnitude, unless hourly workers benefit marginally more,
because of their relatively more deprived endowment of flexibility and need to directly forego
current income to attain it.

Methods
Data and Sample
Data are drawn from the pooled 2002 and 2006 US General Social Survey (GSS), a biannual,
nationally representative, personal interview survey of U.S. households conducted by the
National Opinion Research Center, and its appended Quality of Worklife (QWL) module
(n=4500). The module includes 76 items related to work organization issues including unique
questions regarding working time dimensions. The 2002 sample contains responses from 1,796
employed persons and the identical 2006 another 1,734. Of the entire sample of total employed,
36 percent are salaried, 52 percent are paid by the hour (and the remaining 12% are in

14
See Appendix 3 for a more formal model. Cost savings also may accrue from improved employee health, but
insufficiently to induce all firms to adopt flexible scheduling for all employees (Butler et al., 2009).
15
The association of income with happiness may be relatively stronger for hourly paid workers and at relatively
lower levels of income than at the highest level (see DeVoe & Pfeffer, 2009).
16
For support, see Golden & Wiens-Tuers (2006) and van den Meer & Wielers (2013).
17
In addition, certain jobs or work may provide procedural utility, and thus less disutility, even at long hours.
18
The positive effects may be at least partly negated by the potential jeopardy or backlash perceived when an
employee actually chooses a flexible scheduling option, particularly one that is formally available but is not
supported by supervisors or peers (Galinsky, Bond & Hill, 2005). Bi-directional causality cannot be ruled out here.
independent contract and temporary positions).

Measures
Appendix 1 defines the variables used from the Quality of Worklife (QWL) module data. The
key outcome of interest is contained in the GSS sample—the happiness level item (answered by
n = 2610, 1625 of whom are employed) which asks, “Taken all together, how would you say
things are these days… that you are: very happy, pretty happy, or not too happy?”19 The mean
is 1.8, meaning there are more people in the first than last category. Although it is not known
with certainty that the happiness choice of respondents captures their global happiness rather
than net affect being experienced in the particular moment of the survey taking (see Kahneman
& Krueger, 2006), the preface of the question suggests it is more the former.20
By annual income level, frequency distributions show that in the median income category, 62
percent are “pretty happy,” whereas 32 percent are “very happy” and 8 percent “not too happy.”
In contrast, only 4 percent are “not too happy” and 43 percent are “very happy” in the over $75k,
highest category, while in the lowest (two) income categories, only 26 percent are “very happy”
and over 12 percent are “not too happy.”
Work hours duration is indicated by two questions, the first being the number of hours worked
last week, at all jobs, and the second, by having one or more days per month where extra hours
were worked beyond one’s usual schedule. The average weekly hours worked last week, among
full-time workers in the GSS (2002) sample was 45.9 and 22.6 among part-time workers. The
number of days worked extra per month was 5.5 (5 days among workers with no mandatory
overtime, 9 days among those with mandatory overtime).
Scheduling discretion is measured by three items.21 The first measure, “change start/end time”,
assesses the ability of the respondent to adjust their starting and ending times of work. About 48
percent of the GSS sample feels that they have the ability to often, if not always, adjust their own
starting and ending times, whereas the other half rarely or never do. The distribution of such
perceived daily schedule flexibility is skewed by whether the worker is paid hourly or on salary.
For example 27 percent of salaried but 41 percent of hourly paid workers “never” are allowed to
vary their starting and ending times of work. There are gender differences but only slight—in
hourly jobs, men are slightly more represented in both tails of “never” and “often” in the
frequency of being able to alter their daily start and end times. In contrast, in salaried jobs, men
are more represented than women in being “often” and less so in “never.”
The second measure of schedule discretion, “time off,” reflects the difficulty respondents feel
they have taking time off during the work day to take care of personal or family matters. Almost

19
The GSS contains several questions of this structure, where the responses are excellent, good, fair, etc., or
alternatively, very often, somewhat often, rarely or never.
20
The happiness variable has been criticized by some who feel that it may reflect experiential or ephemeral factors,
which may be even more salient than work-related factors, such as changes in family or social relationships. Also,
the time frame of the question, “these days,” may be interpreted rather openly, so that a respondent might be
considering this day, week or even the past year. In addition, happiness is but one of many potential components in
one’s subjective well-being, albeit a potentially dominant one. Finally, future GSS instruments will move to 7-point
scales, which may allow for greater variation in responses. For the time being, virtually all previous happiness
empirical research using US data have utilized this 3-point scale.
21
The International Social Survey Program also attached a module, Work Orientations III, to the same GSS in 2006,
included a question, “Who sets your work schedule?” For context, this found that about 45 percent in the US
responded that “starting and finishing times are decided by my employer and I cannot change them on my own.”,
whereas only 15 percent felt “entirely free to decide when I start and finish work.” The remaining 40 percent, “can
decide the time … start and finish work, within certain limits.”
three-quarters of the US workforce perceive that it is not very difficult to take time off during the
work day for such matters. Among salaried workers, 53 percent of men and 42 percent of women
consider it “not hard at all” to take time off during the work day. Among hourly workers, the
level is lower for men but slightly higher for women, 47 and 45 percent, respectively. Among
men, 24 percent of salaried and 22 percent of hourly workers find it somewhat or very hard to
take such time off. Among women, the rate is 28 percent for both types of employees.
The third measure of worker discretion is an indicator of whether the respondent reports that they
have the ability to refuse working beyond their usual weekly hours, i.e., that overtime work is
mandatory. Almost 27 percent of workers regard any work beyond their usual weekly hours as
mandatory, required by their employers-- 28 percent of full time and 16 percent of part time
workers. There is virtually no difference between the salaried and hourly work force.
In addition to discretion and duration measures, we include three additional job-related variables
in our regression models. Specifically, the three measures indicate if the respondent works a day
or evening, split or rotating shift, or irregular/on-call shift (day shift is the omitted referent). The
latter may reflect work hours variability. As control variables, we include if a worker is
employed in the non-profit sector or a unionized job. Several demographic control variables
(e.g., age, gender, marital status, number of children) are also included in the model. Table 1
provides descriptive statistics for all variables in the model.22

Table 1: Descriptive Characteristics of the Sample

Salaried Hourly T-test
All Workers N N N
Workers Workers
Demographic Controls
Male (%) 50.8 2,610 51.8 924 53.3 1,378 1.02
Age (Mean, s.d.) 41.9 (13.0) 2,601 43.1 (11.7) 920 40.2 (13.3) 1,374 5.33**
Married (%) 48.2 2,609 55.8 923 42.8 1,378 6.12**
White (%) 76.4 2,610 80.4 924 71.6 1,378 4.84**
The number of a child 1.86+
(Mean, s.d.) 1.57 (1.47) 2,608 1.48 (1.39) 924 1.60 (1.50) 1,376
Survey Year 2002 (%) 34.1 2,610 32.5 924 35.1 1,376 1.28
Income (%) 2,205 819 1,147
less than $20,000 28.4 9.4 40.2
$20,000-$29,999 16.7 11.1 21.9
23.2**
$30,000-$39,999 15.5 17.3 15.3
$40,000-$49,999 12.3 16.4 10.2
over $ 50,000 27.1 45.8 12.4
Work related variables
Working hours (Mean, 41.0 (16.1) 2,596 44.5 (15.0) 918 38.3 (15.3) 1,375 9.50**

22
Appendix 2 illustrates some descriptive bivariate findings related to our key variables--Figure 1a shows that the
ease of taking time off tends to be associated with somewhat greater happiness levels, for both types of workers,
although not always with a perfectly smooth gradation. Figure 1b shows a smoother gradation when examining
happiness levels among the hourly paid, more so than among salaried workers, regarding being able to alter starting
or ending times of work. The pattern of happiness levels by whether or not one’s overtime work is mandatory, in
Figure 1c, appears to be more uniform.
s.d.)
Time off during day(%) 2,600 919 1,375
very hard 11.2 9.7 11.8
somewhat hard 15.6 15.2 16.2 2.03*
not too hard 29.4 30.0 30.6
not at all hard 43.8 45.1 41.4
Change start/end time
(%) 2,597 919 1,374
never 31.5 22.9 41.4
13.4**
rarely 14.6 11.9 18.3
sometimes 19.8 21.2 20.4
often 34.1 44.0 19.9
More days (Mean, s.d.) 5.8 (7.8) 2,563 8.0 (8.6) 908 4.4 (6.7) 1,364 11.1**
Mandatory extra work .51
(%) 26.2 2,574 26.3 917 27.3 1,358
Work schedule (%) 2,602 919 1,378
day shift 73.3 85.0 68.6
X2=
evening shift 10.4 3.5 16.1 111.9**
irregular or oncall shift 9.0 6.6 6.3
split or rotating shift 7.3 4.9 9.0
Belong to Union (%) 14.4 1,749 14.7 631 16.2 912 .79
Work for government or 9.12**
non-profit (%) 24.2 2,603 36.5 924 7.9 1,374
Outcome variable
Happy (%) 2,610 924 1,378
not too happy 10.1 7.1 12.3
6.64**
pretty happy 58.9 54.9 61.6
very happy 31.0 38.0 26.1
Notes: ** p<0.01, * p<0.05, + p<0.1.

Methods and Econometric Analysis: Happiness and Discretion in Work Timing

Econometric analysis is useful in isolating the effect of work schedule discretion on happiness,
holding constant the range of various personal and job characteristics of workers that might
affect happiness uni-directionally. The model estimated is the true frequency of a subjective
well-being outcome given by:

Hj= a + βY1j + β2h + β 3Xj + β 4Flexj+ β 5Zj uj

The dependent variable happiness (Hj) is reported as ordered categories. The independent
variables are the reported income range (Y), measures of the duration of working hours (h), a
vector of control variables (X) including demographics such as gender, marital status, number of
children, age (and age squared), and a vector (Z) capturing other characteristics of their job or
employment arrangement. The focus is on the measured effects of the key independent variables,
which indicate the extent of work schedule discretion, or flexibility (Flex), on the scaled
responses of individuals. Ordinary least squares (OLS) regression models are estimated with the
same set of explanatory and control variables. The constant (a) and also the error term uj account
for other exogenous potential determinants of happiness, for each individual. The size of the
coefficients in the OLS regression results is simply the effect of a unit change in work hours or
degree or frequency of the flexibility indicator reported on the scale of the reported level of
happiness.
In addition to the OLS models, multinomial logistic regression (MLR) estimates are also
conducted for “Very Happy” and “Pretty Happy,” in contrast to “Not Too Happy,” as the omitted
referent group. The estimation results will illustrate the size and statistical significance of the
bivariate that an individual possesses or does not possess a type of flexibility, all else constant,
including the independent effects of income and length of work hours. The regressions all
include controls for the individual’s demographic factors, and some work and related job factors
observable in the GSS. The most basic model is amended sequentially, eventually adding each
flexibility indicator of interest. The last model estimated includes all forms of flexibility
together.

Data Limitations
Several limitations and complications associated with these data must be acknowledged. First,
the indicators are self-rated perceptions which are more prone to potential errors than are
objective measures. In addition, with cross sectional data, indicators of supposed outcomes are
likely to be partly endogenous with the predictor variables, working extra hours and having
discretion. Self-selection may be present and unavoidable if generally happier individuals are
drawn to jobs that feature more flexibility in the timing of work rather than fixed daily work
schedules. Similarly, it is inherently difficult to disentangle the simultaneity if individual self-
reports of happiness are responding to the effects of work conditions or being in less desirable
jobs generally, of which inflexible scheduling is only a reflection.

Results: Work Hours, Worker Scheduling Discretion and Worker Happiness

Table 2 contains the coefficient (β) estimates and z-statistics from the multinomial regressions of
the association between the income gradient and reported happiness level (not too happy relative
to pretty happy and very happy), unadjusted for the control variables and key employment
variables. The middle income category is the referent group. The results are reported for the full
sample, followed by subgroups of respondents who indicate they are paid hourly and those paid a
salary.23 The results show that the highest level of income (above $50k for an individual
respondent) is associated with the greatest happiness, with no other variable controlled. There is
a relatively smooth gradient by income level both above and below the median income group.
The size effect appears highest for the highest income bracket. Importantly, comparing the
results of the multinomial regressions for the split sample, there is a striking difference between
salaried and hourly paid workers and the smoothness of the income gradient disappears for
salaried workers. Salaried workers’ happiness bears no statistically significant association with
higher income. Hourly workers’ happiness, in contrast, is strongly positively associated with
being in the highest income bracket. Specifically, hourly workers in the highest income bracket
are more likely than those in the middle income to report being pretty happy and to report being

23
The “other” subgroup, who are not on a payroll, such as independent contract workers, amount to just over 11
percent of the GSS sample and are omitted in all hourly/salaried subgroup analyses.
very happy (compared to not too happy), whereas these associations are not significant for any of
the other income brackets for hourly workers. 24

Table 2: Multinomial Logistic Regressions, Happiness and Income (unadjusted coefficients) – All
Workers, and by Salaried and Hourly Paid. Unstandardized Coefficients
(Standard Errors in Parentheses)
All workers
Pretty happy Very happy
Coef. z Coef. z
<$20,000 -.045 -.26 -.269 -1.44
(.172) (.187)
$20-29,999 -.012 -.06 -.174 -.79
(.204) (.220)
$40-49,999 .461+ 1.77 .313 1.13
(.260) (.276)
$50,000+ .573** 2.72 .785** 3.60
(.210) (.218)
Observations 2610
Pseudo R2 0.007
Salaried workers
Pretty happy Very happy
Coef. z Coef. z
<$20,000 -.875* −1.97 -.617 −1.36
(.443) (.455)
$20-29,999 .362 .62 .481 .81
(.582) (.593)
$40-49,999 -.142 .33 -.168 -.38
(.427) (.442)
$50,000+ .143 -.43 .030 .09
(.336) (.345)
Observations 924
Pseudo R2 0.008
Hourly workers
Pretty happy Very happy
Coef. z Coef. z

24
For the unreported, “other” classification, the income coefficient estimates and z-statistic significance lie
somewhere in the middle between the hourly and salaried (see, Donnelly, 2010 for possible explanation).
 <$20,000 .268 1.29 −.141 −.61
(.207) (.230)
$20-29,999 .053 .22 −.311 −1.18
(.235) (.265)
$40-49,999 .811* 2.13 .338 .81
(.381) (.417)
$50,000+ .871* 2.29 .892* 2.24
(.380) (.398)
Observations 1378
Pseudo R2 0.008
Notes: ** p<0.01, * p<0.05, + p<0.1
“Not Too Happy” is the omitted category.

To conserve space, the table does not report Prob > χ2 or Log likelihood statistics, which are available
upon request from authors.

Results above combine bottom two categories of income (under $10,000 and $10-20,000) and the top two
categories ($50-75,000 and 75,000+) due to small cell sizes (particularly for the “not too happy”
category for hourly workers in the highest income category and being “very happy” for salaried workers
in the lowest income level . When treated separately, results are replicated but the coefficient is strongest
positive for $75,000 and over, while only marginally significant in the second highest income category.

Table 3 reports the coefficient (β) estimates and z-statistics from the multinomial regressions of
the association between the three key job discretion variables and reported happiness level (not
too happy relative to pretty happy and very happy), adjusted for income and demographic
variables (statistics for income and demographic variables not shown on the table to conserve
space). Each job discretion variable is entered in a separate model, first with “time off during
day”, then with “change start/end time” and then with “mandatory extra work”. Table 3 includes
the coefficients when the models are run on the full sample, as well as when the sample is split
by whether the respondents report being on salary or paid by the hour.
Regarding “time off during day,” Table 3 shows that workers’ subjective well-being benefits
from an ability to take time off from work during the workday to respond to personal or family
matters. The coefficients for both “pretty happy” and “very happy” are statistically significant for
the full sample. In addition, when the sample is split, there is a notably stronger positive
association between the ability to take time off and happiness among hourly than among salaried
workers. Thus, for hourly workers in particular, it appears that flexibility to leave work during a
workday may be a highly valued feature of working time. Of course, it is unclear if this greater
appreciation among hourly workers stems from a generally lower incidence, access or
expectation of such flexibility in hourly paid jobs, a greater need for this kind of flexibility
because of fewer nonwork resources to attend to personal or family matters, or if generally
happier people value such flexibility relatively more than others or sort themselves into jobs that
provide it. Nevertheless, the association is remarkably strong and on the order of moving up
from the median to second highest income bracket in terms of boosting happiness levels.
Table 3 also shows that workers subjective happiness is greater if they report possessing some
degree of control over the timing of their daily start and end times of work. The coefficients are
statistically significant for the sample as a whole, and in particular for hourly workers. As with
“change time,” the coefficients are reduced for the salaried worker subgroup alone. Moreover, it
is evident that the coefficients for control over daily start and end times are somewhat smaller
than those for an ability to take time off during the workday. Again, endogeneity or self-sorting
are a possible reason for these associations that cannot be ruled out. Finally, Table 3 shows no
significant relationship between a lack of control over working extra hours (mandatory overtime,
the third job discretion variable) and happiness. Only for hourly workers, it is found that being
required by an employer to work extra time is negatively related to being “pretty happy” relative
to not happy, although only at a marginally statistically significant level. Thus, the results from
the multinomial logistic regressions reported in Table 3 provide evidence that after controlling
for income and several demographic and work variables, two of the three measures of job
discretion – taking time off during the day to address personal or family matters and discretion
over start and end times – are associated with subjective well-being as measured by the
happiness indicator, especially for workers paid by the hour.

Table 3: Multinomial Logistic Regressions—Happiness and Employee Discretion Over Work Time
(adjusted coefficients) —All workers, Salaried and Hourly Paid
All workers
Pretty happy Very happy
Coef. z Coef. z
Time off during day .199** 3.15 0.372** 5.25
(.063) (.071)
Observations 2590
Pseudo R2 0.060

Change start/end time .131* 2.31 .249** 4.01

(.057) (.062)
Observations 2587
Pseudo R2 0.057

Mandatory extra work -.220 -1.44 -.113 -.68

(.153) (.167)
Observations 2564
Pseudo R2 0.053
Salaried workers
Pretty happy Very happy
Coef. z Coef. z
 Time off during day .147 1.14 .322* 2.35
(.128) (.137)
Observations 916
Pseudo R2 0.0751

Change start/end time .086 .76 .221+ 1.86

(.113) (.119)
Observations 916
Pseudo R2 0.074

Mandatory extra work .075 .24 -.071 -.22

(.314) (.329)
Observations 914
Pseudo R2 0.071
Hourly workers
Pretty happy Very happy
Coef. z Coef. z
Time off during day .289** 3.56 .479** 5.05
(.081) (.095)
Observations 1369
Pseudo R2 0.057

Change start/end time .167* 2.12 .263** 3.01

(.079) (.087)
Observations 1368
Pseudo R2 0.051

Mandatory extra work -.374+ -1.95 -.139 -.65

(.192) (.215)
Observations 1352
Pseudo R2 0.049
Notes: ** p<0.01, * p<0.05, + p<0.1 “Not Too Happy” is the omitted category.

Each separate model was adjusted for the survey year, income, and demographic variables. To conserve
space, the table does not report Prob > χ2 or Log likelihood statistics, which are available upon request
from authors.

Tables 4, 5 and 6 undertake a series of five OLS sequential regressions, treating happiness as a
continuous rather than categorical variable as was presented in Tables 2 and 3. The first model
includes just the income, demographic, and the year dummy variables. Given that the relative
importance of income appeared to be concentrated in the upper income category ($50,000 and
over) in the first set of analyses reported in Table 2, we further distinguish income at the higher
end in these models, splitting the $50,000 and over category into two categories: $50,000 to
$75,000 and $75,000 and over. For hourly workers, however, the single $50,000 and over
category is maintained given low frequency of $75,000 plus incomes among hourly workers.
Then, in three separate regressions (models 2, 3, and 4), two measures of work duration (“hours”
and “more days”) plus each of the three job discretion variables entered separately for each
model are included together with the model 1 variables. Finally, in the full model (model 5), the
three job discretion variables are entered simultaneously, together with three additional work
controls (nonstandard shift timing, union membership25, and whether the job is in the
government or nonprofit sector), and the income, demographic, and work duration variables
previously included. Conducting this series of tests will reveal (1) the extent to which the
association between income and happiness is affected by the inclusion of job discretion variables
(and other work-related variables), and, importantly, (2) whether the association between the key
job discretion variables and happiness that proved important in the multinomial regressions
reported in Table 3 remain important when taking into account the duration of work hours
(models 2, 3, and 4) and a broader set of work variables (model 5). Because model 5 includes
each of the three job discretion variables, it is a particularly strict test of the relationship between
these distinct measures of job discretion and happiness. Table 4 provides results for all workers
combined, Table 5 just for the salaried workers, and Table 6 just for hourly workers.

Table 4: OLS Regressions for Full Sample, Sequential Models 1 through 5 Predicting
Happiness. Unstandardized Coefficients (Standard Errors in Parentheses)
1 2 3 4 5
< $20,000 -0.059+ -0.047 -0.043 -0.042 -0.041
(0.032) (0.033) (0.033) (0.033) (0.033)
$20-29,999 -0.050 -0.049 -0.045 -0.054 -0.053
(0.037) (0.037) (0.037) (0.038) (0.038)
$40-49,999 0.010 0.012 0.013 0.009 0.008
(0.042) (0.042) (0.042) (0.042) (0.042)
$50-74,999 0.042 0.035 0.032 0.031 0.029
(0.038) (0.039) (0.039) (0.039) (0.039)
$75,000+ 0.145** 0.125** 0.109* 0.135** 0.110*
(0.044) (0.045) (0.045) (0.045) (0.045)
Year2002 0.009 0.008 0.009 0.007 0.007
(0.024) (0.024) (0.024) (0.024) (0.024)
Age -0.018** -0.022** -0.022** -0.021** -0.022**
(0.005) (0.005) (0.005) (0.005) (0.006)
Age2 0.000** 0.000** 0.000** 0.000** 0.000**
(0.000) (0.000) (0.000) (0.000) (0.000)
Male -0.008 -0.021 -0.018 -0.011 -0.015
(0.024) (0.024) (0.024) (0.024) (0.025)
White 0.028 0.031 0.020 0.029 0.022
(0.028) (0.028) (0.028) (0.028) (0.028)
Married 0.335** 0.328** 0.330** 0.333** 0.328**
(0.025) (0.025) (0.025) (0.025) (0.025)
One kid -0.106** -0.100** -0.098** -0.105** -0.097**
(0.035) (0.035) (0.036) (0.036) (0.036)
Two kids -0.048 -0.047 -0.048 -0.049 -0.055
(0.034) (0.034) (0.034) (0.034) (0.034)
Three+ kids -0.022 -0.017 -0.013 -0.022 -0.019

25
Missing data on the “union” variable reduces the sample size for Model 5.
 (0.034) (0.034) (0.035) (0.035) (0.035)
Time off during day — 0.065** — — 0.056**
— (0.011) — — (0.012)
Change start/end time — — 0.038** — 0.024*
— — (0.010) — (0.010)
Mandatory extra work — — — -0.001 0.026
— — — (0.027) (0.027)
Hours — 0.002** 0.002* 0.002* 0.002**
— (0.001) (0.001) (0.001) (0.001)
More days — 0.002 0.000 0.001 0.001
— (0.002) (0.002) (0.002) (0.002)
Irreg/Oncall — — — — -0.105*
— — — — (0.043)
Evening — — — — -0.046
— — — — (0.039)
Split/Rotating — — — — 0.010
— — — — (0.045)
Union — — — — -0.015
— — — — (0.041)
Govt/Nonprofit — — — — 0.009
— — — — (0.028)

Observations 2,599 2,537 2,533 2,516 2,505

R-squared 0.087 0.102 0.096 0.090 0.107
Notes: ** p<0.01, * p<0.05, + p<0.1
First, while the pattern of coefficients indicate that happiness levels are positively associated
with higher income (and negatively with lower income), these relationships are statistically
significant only for the $75,000 or over income category (compared to the middle income
category) for the full sample (Table 4) and the $20-$29,999 group for hourly workers (Table 6).
There are no meaningful changes to the income coefficients with the addition of job discretion
and work variables in models 2 through 5.
On the other hand, the “time off during day” and “change start/end time” job discretion variables
continue to show important associations with happiness in the OLS models, as they did in the
earlier multinomial regression models. (The third job discretion variable indicating limited
ability to refuse mandatory work remains insignificant in the OLS models, as was the case in the
earlier models.) In particular, controlling for hours duration, income, and demographic variables,
workers with the ability to take time off during the day to attend to personal and family matters
and workers who can set their own start and end times report greater happiness than their
colleagues without this flexibility. These results hold for the full sample and for salaried and
hourly subgroups. Even in model 5 that includes the three job discretion variables
simultaneously and also includes the additional work controls, the association between taking
time off and happiness remains significant for the full sample and hourly and salaried subgroups,
and the association between changing start and end times and happiness remains significant for
the full sample and marginally significant for the subgroup of workers paid by the hour, but no
longer for salaried workers. These findings indicate that job discretion is an important
component of happiness that future studies on the relationship between income and happiness
might consider. Moreover, these results suggest that taking time off during the day may be an
especially valued form of job discretion for both salaried and hourly workers, as is indicated by
the markedly greater size of the coefficient of this indicator as compared to the coefficient for
“change start/end time” or “mandatory extra work.”
Regarding the duration of weekly work hours, we observe a slight positive association between
working more hours and happiness, and this relationship is significant in the full sample. The
coefficients are largely unchanged for hourly and salaried subgroups but are no longer significant
perhaps because of the smaller sample size. There is no statistically significant relationship found
between working more days and happiness. The lack of negative relationship between work
duration whether measured as hours or days and happiness may in part reflect a “happy worker
effect,” much like the “healthy worker effect,” whereby workers whose health is excellent are
capable of working longer hours.26 Positive mental health might be associated with more hours
devoted to any type of activity, including work. Or, the positive association may be reflective of
something else in the nature of work itself, e.g., more job autonomy, control over pace of work,
stimulating work tasks or other possible sources of process utility.

Table 5: OLS Regressions for Salaried Worker Subsample, Sequential Models 1 through 5
Predicting Happiness. Unstandardized Coefficients (Standard Errors in Parentheses)
1 2 3 4 5
< $20,000 -0.013 0.035 0.021 0.022 0.038
(0.076) (0.077) (0.077) (0.077) (0.077)
$20-29,999 -0.006 0.010 0.005 -0.010 0.016
(0.071) (0.072) (0.072) (0.072) (0.072)
$40-49,999 -0.014 -0.002 -0.003 -0.016 -0.001
(0.062) (0.062) (0.063) (0.063) (0.063)
$50-74,999 -0.016 -0.014 -0.015 -0.017 -0.015
(0.056) (0.056) (0.057) (0.057) (0.057)
$75,000+ 0.062 0.055 0.044 0.052 0.033
(0.059) (0.059) (0.060) (0.060) (0.061)
Year2002 0.002 -0.007 0.002 -0.005 -0.005
(0.041) (0.041) (0.041) (0.042) (0.042)
Age -0.022+ -0.024* -0.026* -0.025* -0.023*
(0.011) (0.012) (0.012) (0.012) (0.012)
Age2 0.000+ 0.000+ 0.000* 0.000+ 0.000+
(0.000) (0.000) (0.000) (0.000) (0.000)
Male 0.048 0.022 0.019 0.037 0.022
(0.040) (0.040) (0.041) (0.041) (0.042)
White 0.044 0.034 0.028 0.033 0.032
(0.049) (0.049) (0.050) (0.050) (0.050)
Married 0.353** 0.346** 0.351** 0.348** 0.335**
(0.043) (0.043) (0.043) (0.044) (0.043)
One kid -0.102+ -0.081 -0.085 -0.088 -0.089

26
Note that models that included workers’ self-reported health were run, and revealed a strongly positive
relationship to happiness, not surprisingly, but are not included herein due to their collinearity with variables such as
work hours duration, age and other demographic variables.
 (0.058) (0.058) (0.059) (0.059) (0.058)
Two kids -0.069 -0.056 -0.060 -0.059 -0.064
(0.054) (0.055) (0.055) (0.055) (0.055)
Three+ kids 0.030 0.042 0.044 0.041 0.033
(0.059) (0.060) (0.060) (0.061) (0.061)
Time off during day — 0.063** 0.055**
— (0.020) (0.021)
Change start/end time — 0.034* 0.014
— (0.016) (0.018)
Mandatory extra work — -0.040 -0.016
— (0.044) (0.045)
Hours — 0.002+ 0.002 0.002 0.002
— (0.001) (0.001) (0.001) (0.001)
More days — 0.004 0.002 0.003 0.003
— (0.002) (0.002) (0.002) (0.002)
Irreg/Oncall — — — — -0.138+
— — — — (0.082)
Evening — — — — -0.162
— — — — (0.107)
Split/Rotating — — — — 0.048
— — — — (0.089)
Union — — — — -0.028
— — — — (0.069)
Govt/Nonprofit — — — — -0.035
— — — — (0.044)

Observations 920 899 898 898 896

R-squared 0.10 0.117 0.110 0.107 0.123
Notes: ** p<0.01, * p<0.05, + p<0.1

Table 6: OLS Regressions for Hourly Paid Worker Subsample, Sequential Models 1 through 5
Predicting Happiness. Unstandardized Coefficients (Standard Errors in Parentheses)
1 2 3 4 5
<$20,000 -0.056 -0.065 -0.054 -0.050 -0.057
(0.041) (0.041) (0.041) (0.042) (0.042)
$20-29,999 -0.080+ -0.088+ -0.082+ -0.091+ -0.097*
(0.047) (0.047) (0.048) (0.048) (0.048)
$40-49,999 -0.019 -0.025 -0.025 -0.022 -0.027
(0.062) (0.062) (0.063) (0.063) (0.063)
$50,000+ 0.064 0.059 0.056 0.046 0.052
(0.059) (0.059) (0.060) (0.060) (0.060)
Year2002 0.037 0.033 0.032 0.031 0.023
(0.033) (0.033) (0.033) (0.034) (0.034)
Age -0.010 -0.013+ -0.013+ -0.011 -0.013+
(0.007) (0.007) (0.007) (0.007) (0.008)
 Age2 0.000 0.000+ 0.000+ 0.000 0.000+
(0.000) (0.000) (0.000) (0.000) (0.000)
Male -0.032 -0.039 -0.029 -0.029 -0.031
(0.034) (0.034) (0.034) (0.035) (0.035)
White 0.029 0.026 0.016 0.023 0.016
(0.037) (0.036) (0.037) (0.037) (0.037)
Married 0.296** 0.297** 0.298** 0.299** 0.300**
(0.035) (0.035) (0.035) (0.035) (0.035)
One kid -0.123* -0.124* -0.113* -0.129** -0.124*
(0.049) (0.049) (0.049) (0.049) (0.050)
Two kids -0.014 -0.025 -0.018 -0.026 -0.035
(0.048) (0.048) (0.048) (0.049) (0.049)
Three+ kids -0.038 -0.048 -0.041 -0.056 -0.051
(0.048) (0.047) (0.048) (0.048) (0.048)
Time off during day — 0.077** 0.068**
— (0.016) (0.017)
Change start/end time — 0.043** 0.026+
— (0.014) (0.015)
Mandatory extra work — 0.021 0.056
— (0.037) (0.038)
Hours — 0.002+ 0.002+ 0.002 0.002+
— (0.001) (0.001) (0.001) (0.001)
More days — -0.002 -0.004 -0.003 -0.003
— (0.002) (0.003) (0.003) (0.003)
Irreg/Oncall — — — — -0.140*
— — — — (0.068)
Evening — — — — -0.001
— — — — (0.045)
Split/Rotating — — — — -0.018
— — — — (0.057)
Union — — — — 0.010
— — — — (0.054)
Govt/Nonprofit — — — — -0.008
— — — — (0.042)

Observations 1,372 1,354 1,353 1,337 1,330

R-squared 0.069 0.087 0.077 0.070 0.093
Notes: ** p<0.01, * p<0.05, + p<0.1

Regarding the other work variables (included in model 5 only), only irregular or on-call work
shows a relationship with happiness. Specifically, workers who report that shift times are
irregular or on-call as compared to daytime work schedules, also report lower happiness. This
supports evidence of the adverse well-being effects on the daily lives of workers who face
variable or unpredictable work hours and schedules (e.g., Henly, Shaefer & Waxman, 2006;
Lambert, 2008). Being a labor union member does not have an association with happiness level;
nor does being employed in government or non-profits.
Some other control variables worthy of note are the pronounced positive effect of being married
with happiness and negative effects of having one (presumably the first) child, especially for
hourly workers. This reinforces findings based on other data elsewhere (e.g., Stevenson &
Wolfers, 2008). Differences by gender, however, are not present. Results for the age control
variable mirror findings regarding the U-shaped pattern of job satisfaction and life satisfaction
(e.g., Clark, Oswald & Warr, 1996; Easterlin, 2006).27

Conclusion, Discussion and Implications

The association of various dimensions of working time and worker self-reported happiness is
revealed herein to have important nuances and implications for working time research. The
General Social Survey (GSS) Quality of Worklife supplements for two years are pooled and
exploited for their unique data generated that distinguishes between hourly paid and salaried
workers, and several types of employee-centered flexibility in scheduling of work. We focused
on three types of work schedule discretion for employees—the ability to vary the starting and
ending times of the workday, ability to take time off during the work day to attend to family or
personal matters and (in)ability to refuse overtime work. In sum, the empirical findings suggest
that discretion over the timing of one’s work matters far more for happiness than does the
duration of working time or income. Indeed, the associations found between job discretion and
happiness are apparently independent of income and work duration, and the associations prove
quite robust over most specifications and control variables. Moreover, the associations of job
discretion with happiness are consistently stronger and more significant among hourly paid
workers. Regarding what kind of job discretion matters, these findings suggest that the ability to
take time off over the course of the day is measurably greater than the ability to vary only the
start or end of the work day. As previously noted, endogeneity cannot be ruled out that happier
people seek or stay in jobs that provide scheduling discretion. Nevertheless, employee input into
working time perhaps yields some procedural utility or hedonic returns that can be quite
important to the employed, particularly to those workers with the least degree of access to it,
hourly-paid workers.
This is consistent with recent findings elsewhere of the effects of more flexible work scheduling
on other potential outcomes of interest for worker subjective well-being--work-life conflict, work
stress and fatigue—which hold even when controlling for duration of work hours and other job
characteristics (Golden et al., 2011). The results lend some support to previous findings that
hourly payment status leads individuals to rely relatively more on using income in the evaluation
of their own happiness, although the income effects are minimized when job discretion variables
are included in the models (DeVoe & Pfeffer, 2009).
Contributions from behavioral economics have been opening the heretofore “black box” of
utility, by examining the economic, and not strictly economic determinants of subjective well-
being and its measurements, such as self-reported happiness scales. Future research should make
the key distinction not only between the effects of the duration of work hours and those from
27
There is growing interest in work hours flexibility over the life course, as a large segment of the work force ages
(e.g., Pitt-Catsouphes & Matz-Costa, 2008; Christensen & Schneider, 2010). When the sample is subdivided by age,
with the control variable for age excluded (above the sample median age of 42, with n= 1242), with results available
upon request from authors. Among older workers, only hourly workers gain happiness with flexible working time,
whereas among younger workers, both hourly and salaried workers gain happiness from flexibility. Relatively older
workers’ happiness is associated with the ability to take off work during the day and to vary one’s start and end
times, no less than it is for the all-age sample.
discretion (control or flexibility) over the timing of work, but how this lack of control
detrimentally affects the character of life and well-being of hourly workers perhaps differently
than salaried workers (Lambert, Haley-Lock, & Henly, 2012; Lambert, 2008; Swanberg et al.,
2011). There seem to be nuanced differences between control over when one does work (starting
and quitting time), does not work, or alternates between work and non-work activities over the
course of a day. Future research also should investigate nuances in the relationship with
happiness levels, hour duration, and control. In unreported results, weekly hours duration and
hours-squared reveal a slight negative effect of hours with a positive exponential, suggesting a
curvilinear effect of longer hours on well-being, which supports Knabe & Ratzel (2010). Since
the ability to take time off during the work day appears to be so crucial, interacting it with
income, shift time and gender might reveal more why it appears to yield greater happiness and
the extent to which it does so differently by income level.
The results show that using a purely economic frame of income levels to explain variation in
well-being indicators would be at best incomplete, thus scheduling autonomy and other job
amenities and working conditions should enter not only worker utility functions but also indexes
comparing well-being across countries. The most pertinent policy implication of the findings is
that while worker well-being is certainly affected by trends in earnings and job opportunities, it
may also be improved by practices and policies that promote more daily discretion over the
timing of workers’ time at work.28 If public policy truly wishes to promote greater gross
domestic happiness, policies that foster the spread in access to and use of autonomy in
scheduling one’s work such as “rights to request” and perhaps to a smaller degree also, “rights to
refuse,” at least for salaried workers, have much promise, and “small necessities laws, which
have been adopted by a handful of US states, might be not so “small” in their power to yield
greater happiness for many of those employed.

References

Aaker, J., Rudd, M., & Mogilner, C. (2011). If money does not make you happy, consider time.
Journal of Consumer Psychology, 21(2), pp. 126–30.
http://dx.doi.org/10.1016/j.jcps.2011.01.004
Ala-Mursula, L., Vahtera, J., Kivimaki, M., Kevin, M.V., & Pentti, J. (2002). Employee control
over working times: associations with subjective health and sickness absences. Journal of
Epidemiology and Community Health, 56(4), pp. 272-79.
http://dx.doi.org/10.1136/jech.56.4.272
Altman, M., & Golden, L. (2007). The economics of flexible work scheduling: Theoretical
advances and paradoxes. In B. A. Rubin (Ed.), Research in the Sociology of Work:
Workplace Temporalities, Vol. 17 (pp. 313–42), Bingley, UK: Emerald Group Publishing
Limited.
Angner, E. (2010). Subjective well-being. The Journal of Socio-Economics, 39(3), pp. 361-68.
http://dx.doi.org/10.1016/j.socec.2009.12.001
Askenazy, P. (2004). Shorter work time, hours flexibility and labor intensification. Eastern
Economic Journal, 30(4), pp. 603-614.
28
Numerous one-time surveys support the potential power of this practice, e.g., “The Key to Happiness at Work:
Schedule Flexibility,” “Minority of Employees Have Flexitime, but Majority Want it” (Tang & MacDermid
Wadsworth, 2010).
Atkinson, C., & Hall, L. (2011). Flexible working and happiness in the NHS. Employee
Relations, 33(2), pp. 88-105.
Bacon, N., Blyton, P., & Dastmalchian, A. (2005). The significance of working time
arrangements accompanying the introduction of teamworking: Evidence from employees.
British Journal of Industrial Relations, 43(4), pp. 681-701.
http://dx.doi.org/10.1111/j.1467-8543.2005.00479.x
Barnett, R. C. (2004, April 29). Work hours as a predictor of stress outcomes. Paper presented at
Long Working Hours, Safety, and Health: Toward a National Research Agenda
conference. Brandeis University, Baltimore MD.
Beckers, D. G. J., van der Linden, D., Smulders, P.G.W., Kompier, M. A. J., Taris, T. W., &
Geurts, S. A. E. (2008). Voluntary or involuntary? Control over overtime and rewards for
overtime in relation to fatigue and work satisfaction. Work and Stress, 22(1), pp. 33–50.
http://dx.doi.org/10.1080/02678370801984927
Berg, P., Kalleberg, A., & Appelbaum, E. (2003). Balancing work and family: The role of high
commitment environments. Industrial Relations, 42(2), pp. 168-188.
http://dx.doi.org/10.1111/1468-232X.00286
Berg, P., Appelbaum, E., Bailey, T., & Kalleberg, A. (2004). Contesting time: International
comparisons of employee control of working time. Industrial and Labor Relations
Review, 57(3), pp. 531-549. http://dx.doi.org/10.2307/4126655
Bielinski, H., Bosch, G., & Wagner, A. (2002). Working time preferences in sixteen European
countries. Dublin: European Foundation for the Improvement of Living and Working
Conditions. Retrieved November 9, 2013 from
http://www.eurofound.europa.eu/publications/htmlfiles/ef0207.htm.
Binswanger, M. (2006). Why does income growth fail to make us happier? Searching for the
treadmills behind the paradox of happiness. Journal of Socio-Economics, 35(2), pp. 366-
381. http://dx.doi.org/10.1016/j.socec.2005.11.040
Booth, A., & Van Ours, J. C. (2008). Job Satisfaction and Family Happiness: The Part-Time
Work Puzzle. The Economic Journal, 118(526), pp. F77–F99.
http://dx.doi.org/10.1111/j.1468-0297.2007.02117.x
Bosch, G. (1999). Working time: Tendencies and emerging issues. International Labour Review.
138(2), pp. 131-149. http://dx.doi.org/10.1111/j.1564-913X.1999.tb00063.x
Boushey, H. (2011). The role of the government in work-family conflict. The Future of Children,
21(2), pp.163-90. Retrieved May 14, 2013, from
http://www.princeton.edu/futureofchildren/publications/docs/21_02_08.pdf
Brenner, M. D., Fairris, D., & Ruser, J. (2004). Flexible work practices and occupational safety
and health: Exploring the relationship between cumulative trauma disorders and
workplace transformation. Industrial Relation, 43(1), pp. 242-266.
http://dx.doi.org/10.1111/j.0019-8676.2004.00325.x
Briscoe, F., Wardell, M., & Sawyer, S. (2011). Membership has its privileges? Contracting and
access to jobs that accommodate work-life needs. Industrial and Labor Relations Review,
64(2), pp. 258-282.
Butler, A., Grzywacz, J., Ettner, S., & Liu, B. (2009). Workplace flexibility, self-reported health,
and health care utilization. Work and Stress, 23(1), pp. 45-59.
http://dx.doi.org/10.1080/02678370902833932
Christensen, K., & Schneider, B. (Eds.)(2010). Workplace Flexibility Realigning 20th-Century
Jobs for a 21st-Century Workforce, Ithaca,NY: Cornell University Press.
Chongvilaivan, A., & Powdthavee, N. (2012). Do different work characteristics have different
distributional impacts on job satisfaction? A study of slope heterogeneity in workers’
 well-being. British Journal of Industrial Relations, Online first, 1-19.
http://dx.doi.org/10.1111/j.1467-8543.2012.00904.x
Clark, A. (2005). Your money or your life: Changing job quality in OECD countries. British
Journal of Industrial Relations, 43(3), pp. 377-400. http://dx.doi.org/10.1111/j.1467-
8543.2005.00361.x
Clark, A., Oswald, A., & Warr, P. (1996). Is job satisfaction U-shaped in age? Journal of
Occupational and Organizational Psychology, 69(1), pp. 57-81.
http://dx.doi.org/10.1111/j.2044-8325.1996.tb00600.x
Clark, A., Frijters, P., & Shields, M. A. (2008). Relative income, happiness and utility: An
explanation for the Easterlin paradox and other puzzles. Journal of Economic Literature,
46(1), pp. 95–144. http://dx.doi.org/10.1257/jel.46.1.95
Cornell University Institute for Workplace Studies. (1999). Overtime and the American Worker.
Ithaca, NY: New York State School of Industrial and Labor Relations.
Robertson, I., & Cooper, G. (2011). Well-being: Productivity and Happiness at Work, UK:
Palgrave Macmillan.
Costa, G., Sartori, S., & Akerstedt, T. (2006). Influence of flexibility and variability of working
hours on health and well-being. Chronobiology International, 23(6), pp. 1125–1137.
http://dx.doi.org/10.1080/07420520601087491
DeVoe, S. E., & Pfeffer, J. (2009). When is happiness about how much you earn? The effect of
hourly payment on the money–happiness connection. Personality and Social Psychology
Bulletin, 35(12), pp. 1602–1618. http://dx.doi.org/10.1177/0146167209346304
Di Tella, R., & MacCulloch, R. (2006). Some uses of happiness data in economics. Journal of
Economic Perspectives, 20(1), pp. 25-46. http://dx.doi.org/10.1257/089533006776526111
Dolan, P., Peasgood, T., & White, M. (2008). Do we really know what makes us happy? A
review of the economic literature on the factors associated with subjective well-being.
Journal of Economic Psychology, 29(1), pp. 94-122.
http://dx.doi.org/10.1016/j.joep.2007.09.001
Dong, X. (2005). Long workhours, work scheduling and work-related injuries among
construction workers in the US. Scandinavian Journal of Work, Environment and Health,
31(5), pp. 329-335. http://dx.doi.org/10.5271/sjweh.915
Donnelly, R. (2010). The organization of working time in the knowledge economy: An insight
into the working time patterns of consultants in the UK and USA. British Journal of
Industrial Relations, 49, pp. s93-s114. http://dx.doi.org/10.1111/j.1467-8543.2010.00826.x
Drago, R., Wooden, M., & Black, D. (2009). Who wants and gets flexibility? Changing work
hours preferences and life events. Industrial and Labor Relations Review, 62(3), pp. 394-
414.
Easterlin, R. A. (2006). Life cycle happiness and its sources: Intersections of psychology,
economics and demography. Journal of Economic Psychology, 27(4), pp. 463–482.
http://dx.doi.org/10.1016/j.joep.2006.05.002
Easterlin, R. A. (2010). Well-being front and center: A note on the Sarkozy report. Population
and Development Review, 36(1), pp. 119–124. http://dx.doi.org/10.1111/j.1728-
4457.2010.00320.x
Felfe, C. (2012). The willingness to pay for job amenities: Evidence from mothers' return to
work. Industrial & Labor Relations Review, 65(2), pp. 427-454.
Friedman, W., & Casner-Lotto, J. (2003). Time is of the essence: New scheduling options for
unionized employees. Work in America Institute and Labor Project for Working Families.
Retrieved November 10, 2007, from http://www.working-
families.org/publications/time_options.pdf
Paul Frijters, Tony Beatton, The mystery of the U-shaped relationship between happiness and
age, Journal of Economic Behavior & Organization, Volume 82, Issues 2–3, May 2012,
Pages 525-542.
Galinsky, E., Bond, J. T., & Hill, E. J. (2005). When work works, Families and Work Institute,
NY: NY.
Gavin, J. H., & Mason, R.O. (2004). The virtuous organization: The value of happiness in the
workplace. Organizational Dynamics, 33(4), pp. 379–392.
http://dx.doi.org/10.1016/j.orgdyn.2004.09.005
Glauber, R. (2011). Limited access: Gender, occupational composition, and flexible work
scheduling. The Sociological Quarterly, 52(3), pp. 472-494.
http://dx.doi.org/10.1111/j.1533-8525.2011.01215.x
Golden, L. (2009). Not formally introduced? Formal vs. informal flexible daily work schedules
among US workers. Industrial Relations, 48(1), pp. 27-54.
http://dx.doi.org/10.1111/j.1468-232X.2008.00544.x
Golden, L., Lambert, S., Henly, J., & Wiens-Tuers, B. (2011). Working time in the employment
relationship: Working time, perceived control and work-life balance. In K. Townsend &
A. Wilkinson (Eds.), Research Handbook on the Future of Work and Employment
Relations (pp. 188-211), Cheltenham UK: Edward Elgar.
http://dx.doi.org/10.4337/9780857936363.00018
Golden, L., & Wiens-Tuers, B. (2006). To your happiness? Extra hours of labor supply and
worker. Journal of Socio-Economics, 35(2), pp. 382-397.
http://dx.doi.org/10.1016/j.socec.2005.11.039
Graham, C. (2009). Happiness around the world: The paradox of happy peasants and miserable
millionaires. Contemporary Sociology, 40(1), pp. 38-9.
Green, F. (2004). Work intensification, discretion and the decline in well-being at work. Eastern
Economic Journal, 30(4), pp. 615-25.
Grönlund, A., & Öun, I. (2010). Rethinking work-family conflict: Dual-earner policies, role
conflict and role expansion in Western Europe. Journal of European Social Policy, 20(3),
pp. 179–95. http://dx.doi.org/10.1177/0958928710364431
Grzywacz, J. G., Carlson, D.S., & Shulkin, S. (2008). Schedule flexibility and stress: Linking
formal flexible arrangements and perceived flexibility to employee health.
Community,Work and Family, 11(2), pp. 199–214.
http://dx.doi.org/10.1080/13668800802024652
Hagerty, M. & Veenhoven, R. (2003). Wealth and happiness revisited - growing national income
does go with greater happiness. Social Indicators Research, 64(1), pp. 1-27.
http://dx.doi.org/10.1023/A:1024790530822
Hamermesh, D. S. (1999). The timing of work over time. Economic Journal, 109(452), pp. 37–
66. http://dx.doi.org/10.1111/1468-0297.00390
Hamermesh, D., & Slemrod, J. (2008). The Economics of Workaholism: We Should Not Have
Worked on This Paper, The B.E. Journal of Economic Analysis & Policy, 8(1), Online
first. http://dx.doi.org/10.2202/1935-1682.1793.
Hamermesh, D., & Lee, J. (2007). Stressed out on four continents: Time crunch or yuppie
kvetch. Review of Economics and Statistics, 89(2), pp. 374-383.
http://dx.doi.org/10.1162/rest.89.2.374
Hancock, E. (2013). Assessing Happiness: How Economic Factors Measure Up. The Park Place
Economist, 21(1/11), pp. 43-57.
Harter J. K., & Arora. R. (2010). The impact of time spent working and job fit on well-being
around the world. In E. Diener, D. Kahneman, & J. F. Helliwell (Eds.), International
Differences in Well-Being (pp. 398-435), New York: Oxford University Press.
Heisz, A., & LaRochelle-Côté, S. (2006). Work hours instability. Perspectives on Labour and
Income, 7(12), pp. 17–20.
Helliwell, J. F. (2003). How's life? Combining individual and national variables to explain
subjective well-being. Economic Modeling, 20(2), pp. 331-
360. http://dx.doi.org/10.1016/S0264-9993(02)00057-3
Helliwell, J. F., Layard, R., & Sachs, J. (Eds.). (2012). World Happiness Report, Columbia
University, NY: Earth Institute.
Henly, J. R., Shaefer, H.L., & Waxman, E. (2006). Nonstandard work schedules: Employer- and
employee-driven flexibility in retail jobs. Social Service Review, 80(4), pp. 609-634.
http://dx.doi.org/10.1086/508478
Heywood, J. S., Siebert, W.S., & Wei, X. (2007). The implicit wage costs of family friendly
work practices. Oxford Economic Papers, 59(2), pp. 275-300.
http://dx.doi.org/10.1093/oep/gpm006
Holzer, H. (2005). Work and Family Life: The Perspective of Employers. In S. M. Bianchi, L.
M. Casper & R. Berkowitz King (Eds.), Work, Family, Health, and Well-being (pp.83-
95), Mahway, N.J.: Lawrence Erlbaum Associates.
Huang, P.H. (2010). Happiness studies and legal policy. Annual Review of Law and Social
Sciences, 6, pp. 405–32. http://dx.doi.org/10.1146/annurev-lawsocsci-102209-152828
Hughes, E. L., & Parkes, K. R. (2007). Work hours and well-being: The roles of work-time
control and work-family interference. Work & Stress, 21(3), pp. 264–278.
http://dx.doi.org/10.1080/02678370701667242
Kahnemann, D., & Krueger, A. B. (2006). Developments in the measurement of subjective well-
being. Journal of Economic Perspectives, 20(1), pp. 3-24.
http://dx.doi.org/10.1257/089533006776526030
Karasek, R. A. (1979). Job demands, job decision latitude, and mental strain: Implications for job
redesign. Administrative Science Quarterly, 24(2), pp. 285-307.
http://dx.doi.org/10.2307/2392498
Kattenbach, R., Demerouti, E., & Nachreiner, F. (2010). Flexible working times: Effects on
employees' exhaustion, work-nonwork conflict and job performance. Career
Development International, 15(3), pp. 279–295.
http://dx.doi.org/10.1108/13620431011053749
Kelliher, C., & Anderson, D. (2010). Doing more with less? Flexible working practices and the
intensification of work. Human Relations, 63(1), pp. 83 -106.
http://dx.doi.org/10.1177/0018726709349199
Kelly, E. L., Moen, P., & Tranby, E. (2011). Changing workplaces to reduce work-family
conflict schedule control in a white-collar organization. American Sociological Review,
76(2), pp. 265-290. http://dx.doi.org/10.1177/0003122411400056
Knabe, A., & Rätzel, S. (2010). Income, happiness, and the disutility of labour. Economics
Letters, 107(1), pp. 77-79. http://dx.doi.org/10.1016/j.econlet.2009.12.032
Kossek, E. E., & Michel, J. S. (2011). Flexible work schedules. In S. Zedeck (Ed.), APA
Handbook of Industrial and Organizational Psychology, Vol. 1 (pp. 535-72).
Washington, DC: American Psychological Association.
Krausz, M., Sagie, A., & Bidermann, Y. (2000). Actual and preferred work schedules and
scheduling control as determinants of job-related attitudes. Journal of Vocational
Behavior, 56(1), pp. 1–11. http://dx.doi.org/10.1006/jvbe.1999.1688
Krueger, A. B. (Ed.). (2009). Measuring the Subjective Well-Being of Nations: National
Accounts of Time Use and Well-Being. Chicago: University of Chicago Press.
http://dx.doi.org/10.7208/chicago/9780226454573.001.0001
Lambert, S. J., Haley-Lock, A., & Henly, J.R. (2012). Schedule flexibility in hourly jobs:
Unanticipated consequences and promising directions. Community, Work and Family
15(3): 293-315.
Lambert, S. J. (2008). Passing the buck: Labor flexibility practices that transfer risk onto hourly
workers. Human Relations, 61(9), pp.1203–1227.
http://dx.doi.org/10.1177/0018726708094910
Lee, B. Y., & DeVoe, S. E. (2012). Flextime and profitability. Industrial Relations, 51(2), pp.
298–316. http://dx.doi.org/10.1111/j.1468-232X.2012.00678.x
Loh, C-P. A. (2009). Physical inactivity and working hour inflexibility: Evidence from a U.S.
sample of older men. Review of Economics of Household, 7(3), pp. 257–281.
http://dx.doi.org/10.1007/s11150-009-9055-8
MacDermid, S. M., & Tang, C. (2009). Flexibility and control: Does one necessarily bring the
other?. In D. R. Crane & E. J. Hill (Eds.), Handbook of Families and Work (pp. 241-264),
Lanham: University Press.
MacKerron, G. (2011). Happiness economics from 35 000 feet. Journal of Economic Surveys,
26(4), pp. 705-735. http://dx.doi.org/10.1111/j.1467-6419.2010.00672.x
Mason, G., & Salverda, W. (2009). Low pay, working conditions, and living standards. In J.
Gautié & J. Schmitt (Eds.), Low Wage Work in the Wealthy World (pp. 35-90), New
York: Russell Sage Foundation.
McCloskey, D. N. (2012). Happyism: The creepy new economics of pleasure. The New
Republic, 243(10), pp. 16-23.
McCrate, E. (2005). Flexible hours, workplace authority and compensating wage differentials in
the US. Feminist Economics, 11(1), pp. 11-39.
http://dx.doi.org/10.1080/1354570042000332588
Messenger, J.C. (2011). Working time trends and developments in Europe. Cambridge Journal
of Economics, 35(2), pp. 295-316. http://dx.doi.org/10.1093/cje/beq022
Mogilner, C., Kamvar, S., & Aaker, J. (2011). The shifting meaning of happiness. Social
Psychological and Personality Science, 2(4), pp. 395-402.
http://dx.doi.org/10.1177/1948550610393987
Noll, H-H. (2011). The Stiglitz-Sen-Fitoussi-report: Old wine in new skins? Views from a social
indicators perspective. Social Indicators Research, 102(1), pp. 111–116.
http://dx.doi.org/10.1007/s11205-010-9738-9
Okulicz-Kozaryn, A. (2011). Europeans work to live and Americans live to work. Journal of
Happiness Studies, 12(2), pp. 225–243. http://dx.doi.org/10.1007/s10902-010-9188-8
Olsen, K. M., & Dahl, S-Å. (2010). Working time: Implications for sickness absence and the
work–family balance. International Journal of Social Welfare, 19(1), pp. 45-53.
http://dx.doi.org/10.1111/j.1468-2397.2008.00619.x
Ortega, J. (2009). Why do employers give discretion? Family versus performance concerns.
Industrial Relations, 48(1), pp. 1–26. http://dx.doi.org/10.1111/j.1468-232X.2008.00543.x
Osberg, L., & Sharpe, A. (2005). How should we measure the ‘economic’ aspects of well being?
Review of Income and Wealth, 51(2), pp. 311–36. http://dx.doi.org/10.1111/j.1475-
4991.2005.00156.x
Pitt-Catsouphes, M., & Matz-Costa, C. (2008). The multi-generational work force: Workplace
flexibility and engagement. Community, Work & Family, 11(2), pp. 215-229.
http://dx.doi.org/10.1080/13668800802021906
Pouwels, B., Siegers, J., Vlasblo, J., & Dirk, J. (2008). Economics letters. 99(1), pp. 72-74.
http://dx.doi.org/10.1016/j.econlet.2007.05.032
Powdthavee, N. (2009). Happiness and well-being. Review for the Beyond Current Horizons
Programme. Beyond Current Horizons. Retrieved January 17, 2013, from
http://www.beyondcurrenthorizons.org.uk/happiness-and-well-being/
Powdthavee, N. (2010). How much does money really matter? Estimating the causal effects of
income on happiness. Empirical Economics, 39(1), pp. 77–92.
http://dx.doi.org/10.1007/s00181-009-0295-5
Reynolds, J. (2003). You can't always get the hours you want: Mismatches between actual and
preferred work hours in the U.S. Social Forces, 81(4), pp. 1171-1199.
Rubery, J., Ward, K., Grimshaw, D., & Beynon, H. (2005). Working time, industrial relations
and the employment relationship. Time & Society, 14(1), pp. 89-111.
http://dx.doi.org/10.1177/0961463X05050300
Schieman, S., & Young, M. (2010). Is there a downside to schedule control for the work-family
interface? Journal of Family Issues, 31(10), pp. 1391-1414.
http://dx.doi.org/10.1177/0192513X10361866
Schubert, C. (2012). Pursuing happiness. Kyklos, 65(2), pp. 245–261.
http://dx.doi.org/10.1111/j.1467-6435.2012.00537.x
Shepard, E., Clifton, T., & Kruse, D. (1996). Flexible work hours and productivity: Some
evidence from the pharmaceutical industry. Industrial Relations, 35(1), pp. 123-139.
http://dx.doi.org/10.1111/j.1468-232X.1996.tb00398.x
Sousa-Poza, A., & Sousa-Poza, A.A. (2000). Well-being at work: a cross-national analysis of the
levels and determinants of job satisfaction, Journal of Socio-economics 29(6), pp.517-
538. http://dx.doi.org/10.1016/S1053-5357(00)00085-8
Sousa-Poza, A., & Henneberger, F. (2002). An empirical analysis of working hours constraints
in twenty-one countries. Review of Social Economy, 60(2), pp. 209–242.
http://dx.doi.org/10.1080/00346760210146235
Spencer, D. (2011). Work is a four-letter word: The economics of work in historical and critical
perspectives. American Journal of Economics and Sociology, 70(3), pp. 563-586.
http://dx.doi.org/10.1111/j.1536-7150.2011.00787.x
Stiglitz J. E., Sen, A., & Fitoussi, J-P. (2009). Report by the Commission on the Measurement of
Economic Performance and Social Progress. Retrieved August 19, 2013, from
http://www.stiglitz-sen-fitoussi.fr/documents/rapport_anglais.pdf
Stevenson, B., & Wolfers, J. (2008). Happiness inequality in the United States. Journal of Legal
Studies, 37(S2), pp. S33-S79. http://dx.doi.org/10.1086/592004
Swanberg, J. E., McKechnie, S., Ojha, M. U., & James, J. (2011). Schedule control, supervisor
support and work engagement: A winning combination for workers in hourly jobs?
Journal of Vocational Behavior, 79(3), pp. 613–624.
http://dx.doi.org/10.1016/j.jvb.2011.04.012
Takahashi, M., Iwasaki, K., Sasaki, T., Kubo, T., Mori, I., & Otsuka, Y. (2011). Worktime
control-dependent reductions in fatigue, sleep problems, and depression. Applied
Ergonomics, 42(2), pp. 244-250. http://dx.doi.org/10.1016/j.apergo.2010.06.006
Tang, C-Y., & MacDermid Wadsworth, S. (2010). Time and workplace flexibility. Families and
Work Institute. Retrieved March 8, 2012, from
http://www.familiesandwork.org/site/research/reports/time_work_flex.pdf
Tijdens, K.G. (2003). Employees' and employers' preferences for working time reduction and
working time differentiation. Acta Sociologica, 46(1), pp. 69-82.
http://dx.doi.org/10.1177/0001699303046001005
Van der Hulst, M., & Geurts, S. (2001). Associations between overtime and psychological health
in high and low reward jobs. Work & Stress, 15(3), pp. 227-40.
http://dx.doi.org/10.1080/026783701110.1080/02678370110066580
Van der Meer, P., & Wielers, R. (2013). What makes workers happy? Applied Economics, 45(3),
pp. 357-368. http://dx.doi.org/10.1080/00036846.2011.602011
Van Emmerik, I.J.H., & Sanders, K. (2005). Mismatch in working hours and affective
commitment: Differential relationships for distinct employee groups. Journal of
Managerial Psychology, 20(8), pp. 712-26. http://dx.doi.org/10.1108/02683940510631462
Van Praag, B. M. S., Frijters, P., & Ferrer-Carbonell, A. (2003). The anatomy of subjective well-
being. Journal of Economic Behavior & Organization, 51(1), pp. 29-49.
http://dx.doi.org/10.1016/S0167-2681(02)00140-3
Virtanen, M., Stansfeld, S.A., Fuhrer, R., Ferrie, J.E., & Kivimaki, M. (2012). Overtime work as
a predictor of major depressive episode: A 5-year follow-up of the Whitehall II study.
PLoS ONE, 7(1), p. e30719. http://dx.doi.org/10.1371/journal.pone.0030719
Wang, J. (2011). Reducing work–life conflict: The role of participating in decision making. In
D. Lewin, B. E. Kaufman, & P. J. Gollan (Eds.), Advances in Industrial and Labor
Relations, Vol. 18 (pp. 153-175), Emerald Group Publishing Limited.
http://dx.doi.org/10.1108/S0742-6186(2011)0000018008
Weeden, K. A. (2005). Is there a flexiglass ceiling? Flexible work arrangements and wages in the
United States. Social Science Research, 34(2), pp. 454-482.
http://dx.doi.org/10.1016/j.ssresearch.2004.04.006
White, M., Hill, S., McGovern, P., Mills, C., & Smeaton, D. (2003). ‘High performance’
management practices, working hours and work-life balance. British Journal of Industrial
Relations, 41(2), pp. 175-195. http://dx.doi.org/10.1111/1467-8543.00268
Weiss, Y. (1996). Synchronization of work schedules. International Economic Review, 37(1), pp.
157–179. http://dx.doi.org/10.2307/2527251
Winder, K. L. (2009). Flexible scheduling and the gender wage gap. B.E. Journal of Economic
Analysis & Policy, 9(1), Online first, Article 30. http://dx.doi.org/10.2202/1935-1682.2197
Wolf, E., & Beblo, M. (2004). Does work time flexibility work? An empirical assessment of the
efficiency effects for German firms. ZEW—Centre for European Economic Research
Discussion paper No. 04-47. Retrieved March 18 2012, from
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=567622.
Wooden, M., Warren, D. & Drago, R. (2009). Working time mismatch and subjective well-
being. British Journal of Industrial Relations, 47(1), pp. 147–179.
http://dx.doi.org/10.1111/j.1467-8543.2008.00705.x
Zuzanek. J. (2012). Does being well-off make us happier? Problems of measurement. Journal of
Happiness Studies, 14(3), pp. 795-815. http://dx.doi.org/10.1007/s10902-012-9356-0

Appendix 1: List of Key Variables

 Dependent Variable
Happy Taken all together, how would you say things are these days-would you say that
you are:
1, very happy, 2. pretty happy, or 3. not too happy?
Income Variable
Reported annual income of individual
1 = "less than $10,000" 2 = "$10,000-$19,999" 3 = "$20,000-$29,999"
4 = "$30,000-$39,999" 5 = "$40,000-$49,999" 6 = "$50,000-$74,999" 7 = "over
$75,000"
Employee Discretion Variables
Change How often are you allowed to change your starting and quitting times on a daily
start/end basis?
time
1 Often 2 Sometimes 3 Rarely 4 Never (REVERSE SCORED)
Time off How hard [difficult] is it to take time off during your work to take care of
during day personal or family matters?
1 Not at all hard; 2 Not too hard; 3 Somewhat hard; 4 Very hard (REVERSE
SCORED)
Mandatory When you work extra hours on your main job, is it mandatory (required by your
extra work employer?)
Hour Duration Variables
Hours If working, full or part time: How many hours did you work last week, at all jobs?
More days How many days per month do you work extra hours beyond your usual schedule?
Paid by a salary or by the hour
In your main job, are you salaried, paid by the hour, or what?

Appendix 2: Descriptive DataFigure 1a, b, c: Cross Tabulations of Happiness with the Three
Indicators of Worker Discretion over Work Schedules, by Hourly and Salary Paid Status, GSS
and QWL data
Appendix 3

Ultimately, it is employers that initially set work schedules (with the possible exception of self-
employment). Employers may offer a traditional, fixed work schedule largely because switching
regimes to provide schedules that better match each employee’s timing preference may not be
entirely cost free.
The degree of scheduling mismatch, on any given shift length experienced by a worker that is
not provided a fully accommodative, flexible schedule is:
* *
( I0 - I0 ) +( In - In )
customized work scheduling system, there will be on-going supervision costs (SC), such as the
administrative and coordination costs of monitoring it once in place, including preventing abuse.
An employer will presumably adopt more scheduling discretion for employees if the added SC in
the short run is exceeded by longer run labor cost savings or improved performance achieved by
better matching actual to preferred schedules (Altman and Golden 2007; Ortega 2009; Lee and
DeVoe 2012). The following equation would transform into monetary terms (via the term, θ) the
net change in labor costs incurred by firms when customized schedules are absent (scaled by the
number of employees involved, N). Thus, a rational firm would:
set I = I * if : SC  | I - I*|
assumes:   1,   0, SC  0
That is, firms would customize work schedules for employees if it provides cost savings that
more than offset the costs of adopting and supervising them. The exponent δ may be simply
equal to one, or perhaps greater if larger deviations from the preferred schedule are harmful to
costs more proportionately than small deviations. In addition, suppose SC is a constant, such that
there are no net economies of scale present by staff size in supervising work schedules.
Introducing the element of time, firms would be induced to implement work scheduling that fits
employees’ preferences if: 1) Employees’ I* changes in ways that would widen the absolute gap
with I; 2) SC decreases, e.g., due to technological innovations in capital, work organization or
human resource management techniques that lower the cost of administering and monitoring
flexible scheduling, such as improved telecommunications devices, diffusion of work scheduling
software tools and self-managed team working; 3) The cost penalty (δ) on the employer for
mismatching increases; 4) The opportunity cost (θ) of reduced retention increases, reflecting,
e.g., a shortage of a specific type of human capital; 5) Either the employer’s time horizon
lengthens, its discounting of future cost savings shrinks or its willingness to risk adopting
flexible scheduling increases.
Therefore, an employer may choose to provide a worker’s desired interval fully, or, perhaps
adjust only partly or gradually. The degree of scheduling flexibility is captured by a coefficient, 
, in the expression,
 I t =  ( I *t - I t -1 ) , 0    1 .
The term  indicates the degree of responsiveness of the actual (or fixed) schedule, I-bar, toward
the preferred I*, when the two deviate. If  is 1.0, a worker has complete scheduling flexibility,
accommodated to adjust to their preferred timing. If it were 0, then the employee works entirely
at the behest of employers’ preferred scheduling.